The Trump Archive collects TV news shows containing debates, speeches, rallies, and other broadcasts related to President-elect Donald Trump. This evolving non-commercial, searchable collection is designed to preserve the historical record for posterity.
The project is a work in progress, with improvements planned for expanding the collection and making searching more efficient.
At launch–blog post here–in January 2017, it is largely hand-curated, with Internet Archive researchers searching out relevant clips, editing them, and placing them in the Trump Archive. To begin, the Internet Archive used video statements fact checked by FactCheck.org, PolitiFact, and The Washington Post’s Fact Checker to seed the archive with relevant material in which there is a strong public interest. Researchers worked to include one copy of each Trump statement, but some duplicates may remain. At launch it contained 700+ TV news shows, 520 hours of video, and more than 500 linked fact checks. The table of fact checks relevant to the Trump Archive can be downloaded as CSV.
We hope to collaborate with researchers to apply machine intelligence to expand this collection and build archives about other public officials representing different political parties. For example, archives in the future may focus on nominees to head federal agencies, members of Congress, Supreme Court nominees and so on.
Such experimentation builds on our experience with first prototyping and then developing the the Political TV Ad Archive. Our first collection of political TV ads, covering ads aired in Philadelphia during the 2014 mid-term elections, was built largely by hand. However, in preparation for the Political TV Ad Archive, we created a new open source tool, the Duplitron, that was able to identify ad airings by deploying audio fingerprinting. During the course of the project, we collected nearly 3,000 ads and documented more than 364,000 ad airings.